First of all, I want to be clear that I am new to AI, although I've learned a bit about classification.

Let's say I have a hierarchical sets of words:

Animal -> 4 legs -> black -> dog
Animal -> 4 legs -> white -> cat                           
Animal -> 8 legs-> black ->spider

and so on...

What is the best algorithm I could use to categorize animals that appear in free text like this:

"I have two **black** **animal** with **4 legs**" 

I want to extract the following: dog

Would it be an SVM even though it's hierarchical?

  • 1
    $\begingroup$ I don't quite follow the impetus behind this. If you have a method to extract & classify "spider" & "cat", & these are perfectly nested w/i higher level categories, why not use that method? Then you could feed the output (s & c) into a simple rule-based algorithm which will add the overarching categories. This seems fairly straightforward to me, although I'm not an AI person. $\endgroup$ – gung - Reinstate Monica Nov 4 '12 at 16:05
  • $\begingroup$ I agree with gung. The question here is whether you completely know the hierarchy in advance or you want to learn it from the data as well. Also I do not see how this is specific to SVM. $\endgroup$ – Bitwise Nov 4 '12 at 18:57
  • $\begingroup$ sorry, i realized my mistake on my way home. i changed the problem description... $\endgroup$ – Ori Price Nov 4 '12 at 20:02
  • $\begingroup$ Not sure if this is what you're looking for, but the Excel 2013 beta can do some neat magic with extracting components of text. They call it "Flash Fill". $\endgroup$ – Jason O. Jensen Nov 4 '12 at 20:07
  • $\begingroup$ @JasonO.Jensen - hey, i want to emmbed this into a web server so i need this as a stand alone feature (or PHP)... anyway i just want a lead on the right algorithm... the code i'll write myself $\endgroup$ – Ori Price Nov 4 '12 at 20:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.